fix(i18n): localize pagination totals across instance and studio pages

[yyqdbngt]

Summary

Replace hard-coded Chinese pagination total texts across the instance and studio pages with localized t() calls, and add the corresponding translation entries:

• common.totalUsers / common.totalGroups / common.totalMessages / common.totalRules interpolation keys (zh + en)
• Applied in consumer.tsx, dlq.tsx (2 spots), message.tsx, topic.tsx, UserManagement.tsx, alerts.tsx
• LiteTopic.tsx: replaced a manual replace('{total}', ...) with the interpolation engine
• Kept alerts.totalRules as the plain stat label ("规则总数"/"Total Rules") and introduced common.totalRules for the paged footer, so the header stat never renders a leftover {count} placeholder

Why

Pagination totals were hard-coded Chinese strings (共 ${n} 个 Group), so the English UI still showed Chinese in the table footers. One change (alerts.totalRules) originally re-purposed a key that is also rendered without arguments as a header stat, which would have leaked a literal {count} into the header; this PR splits the two uses.

Testing

• vitest run src/i18n/__tests__/LangContext.test.tsx src/pages/ops/__tests__/AlertsPage.test.tsx src/pages/studio/__tests__/UserManagement.test.tsx → 3 files, 32 tests passed (interpolation in both languages, no leftover placeholders, alerts stat + paged footer)
• tsc --noEmit → clean
• eslint on changed files → 0 errors (only pre-existing react-refresh/exhaustive-deps warnings)

[RockteMQ-AI]

Summary

This PR successfully localizes pagination totals across the instance and studio pages, replacing hard-coded Chinese strings with proper i18n interpolation. The changes are well-structured with comprehensive test coverage (32 tests), proper separation of concerns, and no breaking changes.

Key improvements:

• Added 4 new translation keys with proper interpolation support
• Fixed a bug where alerts.totalRules was being used for both header stat and pagination footer
• Good test coverage ensures no regression in localization behavior

LGTM — clean implementation with solid test coverage.

---

Automated review by github-manager-bot